Recently, the discussion about whether Cursor can challenge the status of programmers has caused waves in the technical circle. Blogger Feng Yi shared his real experience of using Cursor through a three-day human-machine collaboration experiment, and unveiled the mystery of this hot topic for us.
In Yifeng’s view, Cursor is currently the most powerful product in the field of AI programming. Whether used as an AI programming plug-in, auxiliary tool, or integrated IDE environment, it has demonstrated excellent performance.
Cursors are great at understanding words quickly. Even if the user makes a request in simple vernacular, or even expresses it in Chinese, it can be understood accurately. For example, Cursor can well understand and implement the requirements put forward by Yi Feng in the video and the subsequent optimization requirements, which becomes an intuitive reflection of its ease of use.
Cursor is not an ordinary plug-in, but an integrated development environment or code editor. This gives it an independent voice in file management, operation, testing, debugging, etc. Compared with AI products that are just VS Code or IDA plug-ins, Cursor is not bound by the limitations of the IDE itself. This autonomy provides developers with a smoother programming experience.
The cursor can accurately lock the range to be modified, avoiding the need to regenerate everything based on the prompt word. This function greatly improves programming efficiency and reduces unnecessary duplication of work.
Although Cursor may require user assistance for some customization needs, such as element selectors, the overall responsiveness is satisfactory. After all, different video websites lack unified standards for the definition of likes, coins, and sharing elements. Artificial intelligence has a hard time making decisions when faced with such personalized problems. At this time, it is also reasonable for users to intervene.
It is undoubtedly a wise move for Cursor to choose the integrated development environment route. However, the road has not been easy.
On the one hand, users who are accustomed to using traditional IDEs such as VS Code and Eclipse have accumulated rich user experience, programming interaction methods, debugging experience, and dependent plug-ins and ecological environments during long-term use. These constitute obstacles that are difficult for Cursor to break through in the short term.
On the other hand, if Cursor is just a plug-in like other AI products, then it will be subject to the double limitations of the IDE itself: first, the natural limitations of commercial software technology; second, the natural limitations of commercial software technology. Secondly, when developing AI programming plug-ins, the artificial limitations of the IDE itself, such as not opening internal APIs, hinder functions such as easy file manipulation and submission.
Based on this, Feng Yi believes that the best destination for Cursor may be to be acquired by a company like Jet Brains and join forces to create an integrated development environment with powerful AI programming assistance capabilities.
When it comes to whether artificial intelligence programming will replace programmers, Yi Feng is optimistic and thinks the answer is no. Because many internal requirements require programmers to understand them from a human perspective and then issue disassembly instructions to the AI. The most harmonious way is pair programming between AI and humans. Artificial intelligence can predict human intentions, but the final decision-making power rests with humans.
Unlike autonomous driving, autonomous driving has clear safety goals and can determine accurate routes based on the external environment. AI programming mostly provides multiple solutions that can be selected or requested to be regenerated by humans. Over time, the benefits of AI tools will become clearer to people and allow them to confidently perform tasks they excel at.
In short, AI programming and programmers do not replace each other, but complement each other. As Yifeng said, human-machine collaboration is the best solution for AI programming. What do you think of the future collaboration model between AI programming and programmers? Welcome to share in the comment area, and please like and share so that more people can participate in this technical discussion.
Share on Twitter Share on Facebook
Comments
There are currently no comments